Some small wbe engine doc tweaks
Change-Id: I39c8bb58d910f086b22f4a4591796affc75ea242
This commit is contained in:
@@ -17,8 +17,8 @@ Terminology
|
|||||||
-----------
|
-----------
|
||||||
|
|
||||||
Client
|
Client
|
||||||
Code or program or service that uses this library to define flows and
|
Code or program or service (or user) that uses this library to define
|
||||||
run them via engines.
|
flows and run them via engines.
|
||||||
|
|
||||||
Transport + protocol
|
Transport + protocol
|
||||||
Mechanism (and `protocol`_ on top of that mechanism) used to pass information
|
Mechanism (and `protocol`_ on top of that mechanism) used to pass information
|
||||||
@@ -117,7 +117,7 @@ engine executor in the following manner:
|
|||||||
4. The executor gets the task request confirmation from the worker and the task
|
4. The executor gets the task request confirmation from the worker and the task
|
||||||
request state changes from the ``PENDING`` to the ``RUNNING`` state. Once a
|
request state changes from the ``PENDING`` to the ``RUNNING`` state. Once a
|
||||||
task request is in the ``RUNNING`` state it can't be timed-out (considering
|
task request is in the ``RUNNING`` state it can't be timed-out (considering
|
||||||
that task execution process may take unpredictable time).
|
that the task execution process may take an unpredictable amount of time).
|
||||||
5. The executor gets the task execution result from the worker and passes it
|
5. The executor gets the task execution result from the worker and passes it
|
||||||
back to the executor and worker-based engine to finish task processing (this
|
back to the executor and worker-based engine to finish task processing (this
|
||||||
repeats for subsequent tasks).
|
repeats for subsequent tasks).
|
||||||
@@ -128,7 +128,9 @@ engine executor in the following manner:
|
|||||||
json-serializable (they contain references to tracebacks which are not
|
json-serializable (they contain references to tracebacks which are not
|
||||||
serializable), so they are converted to dicts before sending and converted
|
serializable), so they are converted to dicts before sending and converted
|
||||||
from dicts after receiving on both executor & worker sides (this
|
from dicts after receiving on both executor & worker sides (this
|
||||||
translation is lossy since the traceback won't be fully retained).
|
translation is lossy since the traceback can't be fully retained, due
|
||||||
|
to its contents containing internal interpreter references and
|
||||||
|
details).
|
||||||
|
|
||||||
Protocol
|
Protocol
|
||||||
~~~~~~~~
|
~~~~~~~~
|
||||||
@@ -405,11 +407,11 @@ Limitations
|
|||||||
locally to avoid transport overhead for very simple tasks (currently it will
|
locally to avoid transport overhead for very simple tasks (currently it will
|
||||||
run even lightweight tasks remotely, which may be non-performant).
|
run even lightweight tasks remotely, which may be non-performant).
|
||||||
* Fault detection, currently when a worker acknowledges a task the engine will
|
* Fault detection, currently when a worker acknowledges a task the engine will
|
||||||
wait for the task result indefinitely (a task could take a very long time to
|
wait for the task result indefinitely (a task may take an indeterminate
|
||||||
finish). In the future there needs to be a way to limit the duration of a
|
amount of time to finish). In the future there needs to be a way to limit
|
||||||
remote workers execution (and track there liveness) and possibly spawn
|
the duration of a remote workers execution (and track there liveness) and
|
||||||
the task on a secondary worker if a timeout is reached (aka the first worker
|
possibly spawn the task on a secondary worker if a timeout is reached (aka
|
||||||
has died or has stopped responding).
|
the first worker has died or has stopped responding).
|
||||||
|
|
||||||
Interfaces
|
Interfaces
|
||||||
==========
|
==========
|
||||||
|
|||||||
Reference in New Issue
Block a user